St. Petersburg Paradox Explained: Infinite Expected Value

Quick answer: In the St. Petersburg game, a fair coin is tossed until it lands heads; if the first heads comes on toss k, you win 2^k dollars. The expected value is 1 + 1 + 1 + … = infinity, yet most people would pay only a few dollars to play. Daniel Bernoulli (1738) proposed that the value of money grows more slowly than the amount; applying logarithmic utility to the prize alone, the game is worth about 4 dollars. Real casinos also have finite money, which caps the value.

Here is a game: toss a fair coin until it lands heads. If heads comes on the first toss, you win 2 dollars; on the second, 4; on the third, 8; and so on, doubling each time. How much would you pay to play? Mathematics says the average payout is infinite. Almost nobody would pay more than about 20 dollars. That gap is the St. Petersburg paradox.

The Expected Value Is Infinite

First heads on tossProbabilityPrizeProbability × prize
1\(\frac12\)21
2\(\frac14\)41
3\(\frac18\)81
\(k\)\(\frac{1}{2^k}\)\(2^k\)1

The expected value adds the last column:

\[ E = \sum_{k=1}^{\infty} \frac{1}{2^k} \cdot 2^k = 1 + 1 + 1 + \cdots = \infty. \]

So if you judge purely by expected value, you should pay any finite price to play. That feels absurd: half the time you win just 2 dollars, and three times out of four you win 4 dollars or less.

Bernoulli’s Answer: Diminishing Value of Money

Nicolaus Bernoulli posed the problem in 1713. His cousin Daniel Bernoulli published a resolution in 1738, in the journal of the St. Petersburg Academy, giving the paradox its name. He argued that an extra dollar means less to a rich person than a poor one, so people care about the utility of money, not money itself. With logarithmic utility, \(u(x) = \log x\), the expected utility is finite:

\[ \sum_{k=1}^{\infty} \frac{1}{2^k} \log_2 (2^k) = \sum_{k=1}^{\infty} \frac{k}{2^k} = 2, \]

which corresponds to a sure amount of \(2^2 = 4\) dollars when log utility is applied to the prize alone. If utility depends on total wealth, the fair entry price depends on how rich the player already is. Bernoulli’s idea of expected utility is now a cornerstone of economics, though philosophers still debate whether it settles every version of the paradox.

A Practical Answer: Nobody Has Infinite Money

Any real casino has a limited bankroll. Suppose the most it can pay is \(2^{30} = 1{,}073{,}741{,}824\) dollars (about 1.07 billion), so any larger prize is capped at that amount. Then the expected payout is exactly

\[ \underbrace{1 + 1 + \cdots + 1}_{30} + 1 = 31 \text{ dollars}. \]

(With a cap of exactly 1 billion dollars it is about 30.86 dollars; if the game simply stops after 30 tosses with no prize, it is 30 dollars.)

In this capped model, each doubling of the bankroll adds just 1 dollar of expected value, so even a casino with all the money in the world would push the fair price only to a few dozen dollars.

Other Explanations

  • Risk aversion: most outcomes are small; the huge prizes are vanishingly rare.
  • Ignoring tiny probabilities: people tend to treat a 1-in-a-billion chance as zero.
  • Variance: the game’s spread is infinite, so averages over many plays converge very slowly.

Common Misconceptions

“Infinite expected value means you will win infinite money.” No. Every single game pays a finite prize; the median prize is just 2 to 4 dollars.

“Expected value is useless.” No. It works well when outcomes are bounded and you can repeat many times. The paradox shows its limits when risks are extreme.

Watch: The St. Petersburg Paradox in Video

If you learn better by watching, these videos are worth your time.

The Infinite Money Paradox (Vsauce2). An entertaining look at the game and why nobody pays much to play.
A 300 year old probability paradox | St. Petersburg Paradox (OptWhiz). A clear walk through the expected value calculation.

Check Your Understanding

Five quick questions to see how much of this has stuck. Pick an answer and press Submit; you will see a short explanation either way.

Answer each question and press Submit to see your result right away.

1. What is the probability that the first heads comes on toss 3?
Found a bug? Let us know!
2. What is the expected value of the St. Petersburg game?
Found a bug? Let us know!
3. Who proposed using utility to resolve the paradox?
Found a bug? Let us know!
4. With logarithmic utility applied to the prize alone, what sure amount is the game worth?
Found a bug? Let us know!
5. If prizes are capped at 2^30 dollars, what is the expected payout?
Found a bug? Let us know!

Frequently Asked Questions

What is the St. Petersburg paradox?

A coin game whose expected payout is infinite, yet which most people would pay only a few dollars to play, exposing a limit of judging bets by expected value alone.

Why is the expected value infinite?

If the first heads comes on toss k, the prize is 2^k dollars with probability 1/2^k, so each outcome contributes 1 dollar, and there are infinitely many outcomes.

How did Bernoulli solve the St. Petersburg paradox?

Daniel Bernoulli argued that people value money by its utility, which grows more slowly than the amount. With logarithmic utility the game has a finite value of about 4 dollars.

How much should you pay to play the St. Petersburg game?

In practice only a few dollars to a few dozen dollars, depending on your attitude to risk and on how much money the casino can actually pay.

Keep Exploring

Sources and Further Reading

Related to This Article

What people say about "St. Petersburg Paradox Explained: Infinite Expected Value - Effortless Math"?

No one replied yet.

Leave a Reply